A New Approach for Optimizing Highly Nonlinear Problems Based on the Observer Effect Concept

TL;DR

This paper introduces a novel meta-heuristic algorithm inspired by the observer effect to optimize highly nonlinear, complex engineering problems, demonstrating improved performance in real-world applications like EEG feature learning and generator tuning.

Contribution

The paper presents a new observer effect-based meta-heuristic algorithm that controls memory usage without Tabu-like methods, enhancing optimization in complex nonlinear search spaces.

Findings

Outperforms existing optimizers in complex nonlinear problems

Effective in EEG feature learning and generator parameter tuning

Improved with version updates for better performance

Abstract

A lot of real-world engineering problems represent dynamicity with nests of nonlinearities due to highly complex network of exponential functions or large number of differential equations interacting together. Such search spaces are provided with multiple convex regions peaked with diverse nonlinear slopes and in non-homogenous ways. To find global optima, a new meta-heuristic algorithm is proposed based on Observer Effect concepts for controlling memory usage per localities without pursuing Tabu-like cut-off approaches. Observer effect in physics (or psychology) regards bias in measurement (or perception) due to the interference of instrument (or knowledge). Performance analysis of the proposed algorithms is sought in two real-world engineering applications, i.e., Electroencephalogram feature learning and Distributed Generator parameter tuning, each of which having nonlinearity and complex multi-modal peaks distributions as their characteristics. In addition, the effect of version improvement has been assessed. The performance comparison with other optimizers in the same context suggests that proposed algorithm is useful both solely and in hybrid Gradient Descent settings where problem's search space is nonhomogeneous in terms of local peaks density.